Search has definitely moved beyond traditional text to more niche verticals. The question I'm wondering -- which one of these will truly dominate?
VISUAL SEARCH
What's most interesting in this space here is Riya's recent switch in business model. The CEO, Munjal Shah, originally started Riya because "I found I had 31,246 photos all named DSC0009.jpg." The technology featured advanced facial and text recognition abilities, allowing users to more easily find all the thousands of untitled and untagged photos on their hard drives. In a recent move (Nov 06), they made their switch to focus purely on visual image search (Like.com) after their facial recognition technology failed to be useful to MySpace users. While this has generated a lot of buzz and excitment, not everyone has been happy about the news (including the CEO of Zooomr).
But what's the outcome of this shift? May be too soon to tell, but so far ... positive revenue growth despite decreasing Alexa trends.
Lessons Learned from Riya
- Customer is king.. It takes guts to make a significant product or strategic shift in the company, but at the end, it's important to have a product your customers want. (By the way, I really enjoy Munjal's blog and documentation of his start-up experience. It makes for a great read!)
- You can't make everyone happy. Like.com appeals mostly to (female) shoppers (looking for handbags, jewelry, watches, shoes, and other accessories) and probably won't be as useful for digital photographers that currently use Flickr or Zooomr (looking for artistic shots, scenic, portraits, film, and the like) .
SOCIAL SEARCH
One site I just signed up for is Lijit, a personal network search. Basically, it allows you to search the content of your friends and social network through blogs, photos, bookmarks, online profiles, etc. Check out their about page. Want to try it out?
OTHER SEARCH
For shoppers: Shopping.com, Froogle
For books: Amazon, Half.com
For coders / developers: Krugle
For job seekers: SimplyHired, NimbleCat
For travelers: Kayak, Sidestep
... All this has made me realize how big the web really is. All these different tools out there.
How many people will know about them and -actually- use them? I recently told a friend about Like.com, and she really loved the idea. One thing to say "this is cool." Another to actually use it. Consumers are so stubborn.
MORE RECENTLY
Now, even Wikipedia founder Jimmy Wales is joining in this game with Wikiasari, the people powered search engine.
My Prediction? We will be likely be using multiple search engines for different purposes. I'm a fan of products that focus on one core area and doing it really well. I'm not a fan of the 6-in-1 color printer-fax-phone-copier-scanner device (We have one in our home, and it really sucks.) One thing is for sure: You ( = we = people) will dominate.
VISUAL SEARCH
| Player | Product | Search Type | Description |
| Google's Image Labeler | Tag matching | Random users are asked to tag identical images. If a match is found between these users, that tag has increased relevance during image search. | |
| Yahoo | Flickr | Tags + social networking | Users can tag each photo they upload to Flickr set. They can also search pictures in their social networks (contacts, friends, groups, etc). |
| Zooomr | zooomr.com | Tags + social networking | Like Flickr but larger emphasis on social networking aspect with geotagging features and tracking pictures taken by your friends across the globe. |
| Riya | Like.com | True "visual" search (focused on shopping) | Fell in love with Paris Hilton's new purse? Users can now access similar products online with an easy mouse click .. from a picture! |
What's most interesting in this space here is Riya's recent switch in business model. The CEO, Munjal Shah, originally started Riya because "I found I had 31,246 photos all named DSC0009.jpg." The technology featured advanced facial and text recognition abilities, allowing users to more easily find all the thousands of untitled and untagged photos on their hard drives. In a recent move (Nov 06), they made their switch to focus purely on visual image search (Like.com) after their facial recognition technology failed to be useful to MySpace users. While this has generated a lot of buzz and excitment, not everyone has been happy about the news (including the CEO of Zooomr).
But what's the outcome of this shift? May be too soon to tell, but so far ... positive revenue growth despite decreasing Alexa trends.
Lessons Learned from Riya
- Customer is king.. It takes guts to make a significant product or strategic shift in the company, but at the end, it's important to have a product your customers want. (By the way, I really enjoy Munjal's blog and documentation of his start-up experience. It makes for a great read!)
- You can't make everyone happy. Like.com appeals mostly to (female) shoppers (looking for handbags, jewelry, watches, shoes, and other accessories) and probably won't be as useful for digital photographers that currently use Flickr or Zooomr (looking for artistic shots, scenic, portraits, film, and the like) .
SOCIAL SEARCH
One site I just signed up for is Lijit, a personal network search. Basically, it allows you to search the content of your friends and social network through blogs, photos, bookmarks, online profiles, etc. Check out their about page. Want to try it out?
OTHER SEARCH
For shoppers: Shopping.com, Froogle
For books: Amazon, Half.com
For coders / developers: Krugle
For job seekers: SimplyHired, NimbleCat
For travelers: Kayak, Sidestep
... All this has made me realize how big the web really is. All these different tools out there.
How many people will know about them and -actually- use them? I recently told a friend about Like.com, and she really loved the idea. One thing to say "this is cool." Another to actually use it. Consumers are so stubborn.
MORE RECENTLY
Now, even Wikipedia founder Jimmy Wales is joining in this game with Wikiasari, the people powered search engine.
My Prediction? We will be likely be using multiple search engines for different purposes. I'm a fan of products that focus on one core area and doing it really well. I'm not a fan of the 6-in-1 color printer-fax-phone-copier-scanner device (We have one in our home, and it really sucks.) One thing is for sure: You ( = we = people) will dominate.
technorati tags: future of search, visual search, like.com, flickr, zooomr, , ligit, google, yahoo, image search, vertical search, social search, personal network search, wikiasari
Comments (6)
So out of curiosity, what's your take on Yahoo!'s push towards search and application integration? Microsoft's attempt to integrate services into the Live platform?
It seems to me that there's a reason why there's so much consolidation going on, with Google now buying major competitors (i.e., YouTube) and Yahoo! buying multiple niche players. It seems as though the big guys are trying to take vertical search and consolidate, no?
Posted by Anonymous | December 27, 2006 11:03 PM
Anonymous,
Yes, I do agree that is a trend. I should probably clarify my statement of "we will likely be using multiple search engines." What I really meant was vertical search engines do a better job at their specific niche compared to a general search engine. For instance, a few friends of mine built a search engine algorithm to do searches within a local domain (e.g. www.upenn.edu) for their senior design project in computer science. It turned out to display more relevant results than what Google provided for the same keyword searches.
That being said, I think that consumers like (pardon the use of the word) "platform" products - going to one place for core needs. So, even as this consolidation happens, the underlying technology is still better than what currently exists.
Posted by Jing | December 27, 2006 11:28 PM
With regular search engines, there really seems to be a trade-off between relevance (a euphamism for quality, really) and quantity. Domain-specific searches in particular suffer from this tradeoff, but search engines in general are very susceptible to this problem.
I remember reading a paper that indicated that Google is capable of presenting something like .2% of the entire WWW. Granted, I'm pretty sure these numbers include each indvidual page of each web-site set, but that's still a slim number.
Restricting indexing with more specific constraints (i.e., vertical search) actually seems to ameliorate both problems if the user knows what they want.
So, for example, a shopping search will index only those items suggested to be products (sometimes specific products). Although this brings additional problems (the quantity problem: assuming vertical search engines index conservatively, they are likely to miss multiple products), this seems to be an appropriate way to maximize the number of results indexed as well as the relevance of such indexing.
Just thinking out loud, but I like the post. Thanks!
Posted by Anonymous | December 28, 2006 2:03 AM
One of the best social search engines and visual search has been available since the beginning of web 2.0: Del.icio.us. Any time I'm looking for interesting images (whether for a project or my blog) I always check out del.icio.us.
The best proof of del.icio.us's power is in generic searches. Search Google for "funny images" and you get the worst trash links imaginable. Do the same search for "funny images" on del.icio.us and you have results of 20x quality over Google.
Just another possibility for your list.
Posted by Andrew | December 28, 2006 3:31 AM
Thanks Andrew! I tried that query and it worked wonders.
Posted by Jing | December 30, 2006 8:28 AM
Have you Tried Trabber? it is a new travel search engine that you might also like. Here is the address: http://www.trabber.com
Ciao
Posted by dan | January 7, 2007 4:32 AM